Every operant contingency has a source. Sometimes another person delivers or removes the consequence; sometimes the behavior produces the consequence all by itself. Knowing which one you are dealing with tells you where to look for the reinforcer and which interventions have a chance of working.
Why it matters
Treatment follows function. If a child's hand flapping is maintained by attention from adults, changing how adults respond can reduce it. If the same hand flapping is maintained by the visual or tactile stimulation it creates, adults can ignore it all day and nothing changes, because their behavior was never part of the contingency.
In case work and practice questions, this distinction usually shows up inside a vignette or a set of functional analysis results. You need to identify where the consequence came from, decide whether it was added or removed, and pick an intervention that actually touches the maintaining reinforcer.
Key ideas
- The question is "who or what produced the consequence?" If the consequence depends on another person's behavior, the contingency is socially mediated. If the response produces the consequence directly, with no one else involved, it is automatic.
- Both kinds can be positive or negative. Humming that produces a sound the person enjoys is automatic positive reinforcement. Scratching that relieves an itch is automatic negative reinforcement. A caregiver handing over a tablet after a request is socially mediated positive reinforcement. A teacher removing a worksheet after screaming is socially mediated negative reinforcement.
- Punishment can be automatic, too. Touching a hot pan produces pain without anyone else acting, and touching hot pans becomes less likely. A coach's reprimand that reduces a behavior is socially mediated punishment.
- Social mediation does not require intent. Peers who laugh at a classmate's jokes are mediating the consequence even if no one planned it as a reinforcer.
- Form does not tell you function. Rocking, mouthing, head hitting, and vocal noises can be automatically reinforced for one person and socially mediated for another. Only an analysis of the contingency, ideally with data, tells you which.
- "Automatic" is not the same as "unlearned." The term refers to how the consequence is delivered, not to whether the reinforcer is unconditioned or conditioned, and not to whether the behavior is innate.
Definitions
- Automatic reinforcement: reinforcement in which the response itself produces the reinforcing consequence, independent of anyone else delivering it.
- Socially mediated reinforcement: reinforcement in which the consequence is delivered or removed through the behavior of another person, such as attention, access to items, or the end of a demand.
- Automatic punishment: punishment in which the response directly produces the consequence that makes it less likely in the future.
- Sensory extinction: an extinction procedure for automatically reinforced behavior that masks or removes the sensory consequence the response produces.
Examples from a session
Sorting the source of the consequence
- Automatic positive: Mei, a 5-year-old in a clinic, lines up blocks and sights along them. She does this at the same rate whether or not anyone is in the room. The visual pattern the response creates is the likely reinforcer.
- Automatic negative: Mr. Alvarez, 68, rubs his healing knee incision, which briefly eases the itch. No one else is involved, and rubbing persists.
- Socially mediated positive: In a middle school classroom, Darius calls out answers and his teacher responds to each one. Calling out increases.
- Socially mediated negative: At home, Lena, 7, cries during bath time, and her dad ends the bath early. Crying during baths increases.
Functional analysis data often point to the source. Responding that stays high when the person is alone and without materials, or that is high across every condition with no clear pattern, suggests automatic reinforcement. Responding that is high mainly in the attention, tangible, or demand condition suggests a socially mediated function.
Common mistakes / exam traps
- Assuming repetitive behavior is automatic. Stereotypy is a description of form. Check the data before calling the function automatic.
- Calling a contingency automatic because no adult acted. Peers, siblings, and strangers can all mediate consequences.
- Calling a contingency socially mediated because someone set up the setting. A person who installs a device or rewires a switch acts once; if each consequence then follows the response with no one acting, the contingency is automatic.
- Equating "automatic" with "positive." Automatic reinforcement can remove stimulation as well as add it.
- Mixing up the source with the type of reinforcer. Food delivered by a parent is an unconditioned reinforcer and socially mediated at the same time. The two classifications answer different questions.
Exam trap
Planned ignoring vs. sensory extinction: both withhold a consequence, and the difference is which one. Planned ignoring withholds attention, so it works as extinction only when another person's attention is the reinforcer. Sensory extinction masks or removes the stimulation the response produces by itself. When rates stay similar with or without people present, ignoring leaves the reinforcer intact; look instead at sensory extinction, matched or competing stimulation, or an alternative response that produces similar stimulation.
Exam trap
Automatic vs. socially mediated is a separate question from positive vs. negative, so answer them one at a time. Work in order: did the behavior increase or decrease (reinforcement or punishment)? Was a stimulus added or removed (positive or negative)? Did another person's behavior produce that change (socially mediated) or did the response produce it directly (automatic)?
Quick check
During recess, Kai, age 9, makes loud animal noises. His classmates laugh each time, although no one means it as a reward. Over the month, the noises increase. Which contingency best describes this?
Related topics
- Positive vs. negative reinforcement (B.4)
- Positive vs. negative punishment (B.5)
- Kinds of reinforcers: unconditioned, conditioned, and generalized (B.7)
- Operant vs. respondent extinction, as procedure and as process (B.11)
- Motivating operations in action (B.16)
- Designing and evaluating functional analyses (F.6)